Home Cellular health A miniature fluorescence microscope for multiplane imaging

A miniature fluorescence microscope for multiplane imaging


All experimental procedures and animal care were reviewed and approved by the Institutional Animal Care and Use Committee (IACUC), Intramural Research Program, National Institute on Substance Abuse, National Institutes of health. All animal studies were performed according to the guidelines of the IACUC approved protocol and in accordance with ARRIVE guidelines.

Viral injection

The viral injection procedure was similar to that previously described27, and briefly described here. Six male C57BL/6 J mice (3-4 months old, body weight ~25 g) were injected with AAV-pgk-Cre (Addgene, #24593-AAVrg) into Nucleus Accumbens and pGP-AAV-syn- FLEX-jGCaMP7f-WPRE (Addgene, #104492-AAV1) in the prelimbic cortex (PrL). Briefly, mice were anesthetized with 2% isoflurane in oxygen at a flow rate of 0.4 L/min and mounted on a stereotactic frame (Model 962, David Kopf Instruments), while temperature was maintained at 37°C using a temperature monitoring system (TCAT-2DF, Physitemp). Sterile lubricating eye ointment (Dechra Veterinary Products) was applied to the corneas of mice to prevent drying. A hole was drilled through the right side of the skull above the injection site (A/P: + 1.9 mm; M/L: − 0.3 mm) using a round bur of 0.5 mm in diameter on a high-speed rotary micro drill (19007 -05, Fine Science Tools). A total of 500 nl of virus (a titer of 6.75e12 GC/mL) was injected using stereotactic coordinates (A/P: +1.9 mm, M/L: −0.3 mm, D/V : − 1.7 mm, 0° angle) at a flow rate of 25 nl/min with a micro pump and a Micro4 controller (World Precision Instruments). After injection, the injection needle was held in the parenchyma for 5 min before being slowly withdrawn. The hole on the skull was then sealed with bone wax and the skin was sutured. After surgery, Neosporin ointment was applied to the closed skin incision line. Mice were injected subcutaneously with buprenorphine (0.05 mg/kg) and returned to their home cage to recover from anesthesia in a 37°C isothermal chamber (Lyon Technologies, Inc). Mice were maintained on ibuprofen (30 mg/mL in water) ad libitum for at least 3 days after surgery.

Implantation of gradient index lenses (GRIN)

The procedure for implanting the GRIN lens was similar to that previously described27, and briefly described here. One week after viral injection, a 1 mm diameter gradient index (GRIN) lens (GRINTECH GmBH) was implanted into the mouse brain in the mPFC. Briefly, mice were anesthetized with ketamine/xylazine (ketamine: 100 mg/kg, xylazine: 15 mg/kg) and a 1 mm diameter craniotomy was generated in the right hemisphere above the coordinates (A/P: +1.9mm, M/L: −0.7mm). Freshly prepared artificial cerebrospinal fluid (aCSF) was continuously applied to exposed tissue throughout surgery to prevent dehydration of brain tissue. Brain tissue above the PrL, along a direction of a 10° angle offset laterally to a depth of 1.8 mm, was precisely removed using vacuum aspiration through a 30-gauge blunt needle attached to a custom-built three-axis motorized stereotaxic device modified from a commercial stereotaxic frame (Model 962, David Kopf Instruments). Once the brain tissue above the mPFC was removed and the surgical site was free of blood, a GRIN lens was slowly lowered into the mPFC and secured to the skull using dental cement (DuraLay). Approximately one month after implantation of the GRIN lens, a custom baseplate was mounted on the mouse’s head with dental cement. The miniscope was then fixed to the base plate using 3 screws.

miniscope design

The miniscope consists of the main body, designed in Solidworks and 3D printed in black resin (Protolabs, Maple Plain, MN), light source (470 nm high power LED XPEBBL-L1-0000-00302, Cree LED), optics (see Table 1) and a CMOS sensor (MT9V022, Aptina/Onsemi) mounted on a custom PCB that connects to the data acquisition system as described in28. A schematic of the mechanical design of the miniscope is shown in Fig. 1a. We used a varifocal liquid lens (A-16F0-P12, Corning Varioptic), which allows rapid change of focal length, controlled and synchronized with the miniscope’s image sensor via the custom FPGA control board.

Table 1 Nomenclature.
Figure 1

Characterization with a miniscope. (a) Schematic of the miniscope design. The figures in blue correspond to the headings of Table 1. (b) Top: Image of light reflected off a 1951 3″ × 3″ USAF target. Inset: Enlarged detail of Group 7, Element 4 (181 lp/mm). Bottom: Intensity values ​​for highlighted pixels in the inset row (left) and column (right), spanning two row pairs over 11.05 μm, and resulting in a magnification factor of 3 x with a resolution of 5.52 μm.

Comparative analysis

We tested the optical performance of the miniscope by imaging a 1951 USAF target (Fig. 1b), measuring a vertical and horizontal resolution of 5.52 μm for a magnification factor of approximately 3×.

Axial scan performance was tested by positioning the miniscope vertically (Fig. 2a) on a 45° depth of field target (Edmund Optics DOF ​​5–15, Fig. 2b) and recording the horizontal lines of the 15 lp/mm section (Fig. 2c) at different plane depths, ranging from 0 to 100% of the PWM duty cycle of the 26-step liquid lens drive signal (Fig. 2d,e). For each depth of shot, we calculated the average of a sequence of 30 images. Changing the drive signal produced a change in focal plane depth which was estimated using the line averaging curve fitting parameters of the images: averaging along y for each duty cycle stage, the resulting curves were approximated as a Fourier series truncated at the first harmonic (tuned to the frequency of the line pairs, ω = 66.67), multiplied by an exponential envelope curve, of which the central parameter b indicated the estimated depth of the focal plane:

Figure 2
Figure 2

Assessment of focal plane change. (a) The assembled miniscope was placed vertically on the depth of field target to evaluate its axial scanning performance. (b) A photo of the target at 45° depth of field. (vs) Detail of the depth of the deposited target, highlighting a detail of the imaged field of view from the 15 lp/mm section (inset) (D) Images captured at 3 different focal plane depths on a 45° depth of field target with 15 lp/mm. (e) Reconstructed field of view (FOV), merging the 40 lines around the estimated focal length for each depth. (F) Left: average of rows of images captured at different focal plane depths, spanning a PWM cycle from 0 to 100% of the liquid lens drive signal. The green dots indicate the estimated focal plane depth for each frame, and the red line is their linear fit (5.98 μm slope per percent change in PWM duty cycle). The dotted white lines represent the 3 frames in (D). Right: Average of pixel rows for the image captured at 66% PWM duty cycle control signal and displayed as magenta dots in the left panel. The center of the fitted exponential envelope curve (green dotted line) is used to estimate the depth of the focal plane.

$$acdot {e}^{-{left(frac{xb}{c}right)}^{2}}cdot ({A}_{0}+{A}_{1} mathrm{cos}left(omega xright)+{B}_{1}mathrm{sin}(omega x))$$

The focal plane change as a function of PWM duty cycle of the liquid lens could be approximated linearly (R2= 0.99287) with a slope of 5.98 μm per percent change in duty cycle (Fig. 2f).

To assess whether the image could stabilize over time between two frames, we recorded 200 frames at 10 fps, alternating focal plane depth at each frame, for 5 sessions (Fig. 3a).

picture 3
picture 3

Fast focal plane switching. (a) Focal plane depth (estimated using the linear fit in Fig. 2c) for 5 recorded sessions alternating between two focal planes at each frame, with 100 frames recorded at each depth. (b) RMS average pixel intensity for the image difference between the average of 100 images recorded at each of the alternating plane depths (red and blue for plane 1 and plane 2, respectively) and the statically recorded images over all values PWM cycle control. The dotted lines mark the PWM cycle used on each of the two planes. The local minima of the RMS difference curve denote a greater similarity between the two images.

We then calculated the RMS difference between the average of 100 images for each plane and the static images captured over the 26 steps of the PWM cycle (Fig. 3b): for all sessions, over all focal plane differences, the local minima of the RMS curve (and the greatest similarity between the images) coincided with the focal plane at which the alternate planes were recorded, indicating a close match between the static images captured at each depth and the corresponding alternate images.